Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add gke-labs/kube-agents --skill version-controlgit clone --depth 1 https://github.com/gke-labs/kube-agentsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/gke-labs/kube-agents/version-control)<a href="https://agentmods.dev/skills/gke-labs/kube-agents/version-control"><img src="https://agentmods.dev/badge/skills/gke-labs/kube-agents/version-control/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/gke-labs/kube-agents/version-control"><img src="https://agentmods.dev/badge/skills/gke-labs/kube-agents/version-control.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00065 | $0.05011 |
| Opus 5.5 | $0.00026 | $0.02004 |
| Sonnet 5.5 | $0.00013 | $0.01002 |
| Haiku 4.5 | $0.00006 | $0.00501 |
Grade A, and why
version-control scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 321 lines — stays where its author put it; the contents beside it link to each section on GitHub.
version-control - remote operations through vcs.py, local operations with git
Version control here is abstracted. Everything is one of two things, and getting an operation into the right half is most of using this skill.
Remote operations — anything that crosses the network or spends a
credential. These go through scripts/vcs.py, which speaks to a broker in
another container. Only the broker holds the credential; nothing here does.
The remote verbs are exactly these:
capabilities, identity, clone, publish,
proposal create|list|view|comment|update|close|commits|acknowledge,
issue create|list|view|comment|update|close, label ensure,
remote-branch view|delete.
update edits a title or body and adds or removes labels; close closes;
commits lists the revisions on a proposal's source branch; acknowledge
reacts to a comment (its id and kind come from view --comments) so its
author sees it was read — capabilities says whether this forge supports it.
issue list --query <text> searches, and --without-labels skips issues
carrying any of them — that is how a queue is read for unclaimed work.
label ensure creates a label or updates it if it exists. identity says
who this install is on the forge and, with --login, whether that login may
write to the repository. remote-branch view says whether the shared
repository still holds a branch and at which revision; remote-branch delete
removes one only when it is spent — under platform-agent/, no open proposal
on it, and its tip exactly what a merged or closed proposal this install opened
from this repository carried — and only
at the --revision you just read.
Local operations — everything else. clone unpacks a real working copy
onto this filesystem and prints its path. Inside it, use git —
git log, git show, git blame, git grep, git diff, git status,
git branch, git commit, git ls-files --stage — and every read works as you
expect. Read files in the working copy with cat, rg, or anything else. You
do not need vcs.py for any of this and it is faster without it.
This git holds no credential and cannot reach a forge: its HTTP transport
helpers are not in the image, so an https:// URL fails with
'remote-https' is not a git command, and there is no ssh client, so an
ssh:// URL fails with cannot run ssh. Seeing either message means you asked
it for the one thing it does not do; the answer is a vcs.py verb. There is no
gh here either, for the same reason.
vcs.py also offers log, show, annotate, files, grep, diff,
status, branch and commit as thin wrappers over that same local git, for
when you want their JSON output. They make no network call. Use whichever you
prefer; the wrapper is a convenience, not the sanctioned path.
The reason for the split is that the history is here rather than there. clone
brings a repository down as a git bundle, so the full object graph — every
commit, every parent, every tree entry with its mode — is on this disk. Your
revisions go up the same way, as a bundle, which is why a branch of five
commits arrives as five commits. Nothing that came out of the repository is
ever executed beside the credential.
The script is /opt/data/profiles/platform/skills/version-control/scripts/vcs.py (scripts/vcs.py inside the skill directory). Every subcommand prints one JSON object on
stdout. --repo takes owner/name or a full URL, and the broker decides which
forge that is. On an install that serves more than one forge, owner/name is
refused: name the repository by its URL.
Verb names are the version-control concept; the spelling you know is an alias.
annotate/blame, log/history, files/manifest, grep/search,
publish/push, proposal/pr/mr, create/open all work.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago Changed 76acfe5208aa
- 8d ago Changed · +6 lines 641833918122
- 14d ago Changed · +25 lines ed55d2edebba
- 15d ago Changed · +1 lines 8f197e9885e8
- 18d ago Changed · -9 lines 5535ee1cc7e0
- 20d ago Changed · -2 lines 8a6b376f21d3
- 21d ago Changed · +26 lines 6dc34ec1c499
- 23d ago Changed · +55 lines 5dfa7fbac293
- 1mo ago First seen · 219 lines · 65 tokens per session scan A d196c23a9c29
version-control is a skill published in the GitHub repository gke-labs/kube-agents (109 stars, last pushed today), licensed Apache-2.0. It adds 65 tokens to every session and 5,011 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-16.
Other skills, from other repositories
create-pr
Creates a GitHub PR with a Linear-ticket-prefixed title and a decision-led, narrative description for Prisma 8. Use when the user wants to create a pull request, open a PR, or submit changes for review.
git-commit
Generate well-formatted git commit messages following conventional commit standards.
repo-sync
Manage automatic Git syncing on macOS with repo-sync. Use only when the user explicitly invokes /repo-sync.
mcore-split-pr
Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.
contributing
How to contribute to evlog, covering commit and PR conventions, changesets, the Definition of Done, testing rules, and the authored skills that walk through building a new adapter, enricher, framework integration, or map rule. Load this for any question about contributing, opening a PR, or adding something to the…
engineer-bot-ops
A workflow for carrying an engineering task from scope through implementation and reviewable delivery. It works with a repository's branch, existing changes, rules, checks, and permissions while separating completed, unverified, and next-step work.